introduction and overview of redfish - snia | … fujitsu, huawei, ibm, insyde software, mellanox,...

2016 Storage Developer Conference. © Dell. All Rights Reserved.

Dell - Internal Use - Confidential

Introduction and Overview of Redfish

Patrick Boyd Dell




Agenda

Who is behind Redfish? What is Redfish? Simple Example Why HTTPS, REST, and JSON Slightly more complex example Discovery Creation example

2



Who is behind Redfish?

3

Supporting Companies

•AMI, Fujitsu, Huawei, IBM, Insyde Software, Mellanox, Microsemi, NetApp, Oracle, OSIsoft LLC, Quanta Computer, Seagate, Western Digital

Industry Alliance Partners

•OpenCompute Project •UEFI - Collaborating on Firmware Update and Host Interface work •SNIA – Collaborating on Storage modeling / alignment between SSM and Redfish •TGG – Pursuing relationship to work on Power/Cooling (existing DMTF Alliance Partner)

Promoter Companies



What is Redfish?

Industry Standard RESTful API for IT Infrastructure HTTPS in JSON format based on Odata v4 Equally useable by Apps, GUIs, and Scripts Schema-backed but human-readable

4



What is Redfish?

First release focused on Servers A secure, multi-node capable replacement for

IPMI-over-LAN Add devices over time to cover customer use

cases & technology Direct attach storage, PCIe and SAS switching,

NVDIMMs, Multifuction Adapters, Composability Intended to meet OCP Remote Machine

Management Requirements 5



What is Redfish?

Expand scope over time to rest of IT infrastructure Working with SNIA to cover more advanced

storage. See “Overview of Swordfish” and “Swordfish Deep-

dive” later in the conference Plan on working with partners like Green Grid

to cover Power/Cooling Goal is to accommodate or map existing

network switch standards over time 6



Simple Example

7

Example Python code to retrieve serial number from a server:

Output is:

*Example uses Redfish ComputerSystem resource, Authentication not shown

rawData = urllib.urlopen(‘http://192.168.1.135/redfish/v1/Systems/1’) jsonData = json.loads(rawData) print( jsonData[‘SerialNumber’] )

1A87CA442K



Simple Example

8

From the Linux command line

Output is:

*Example uses Redfish ComputerSystem resource, Authentication not shown **jq is a command line JSON parser available here https://stedolan.github.io/jq/

curl http://192.168.1.135/redfish/v1/Systems/1 | jq ‘.SerialNumber’

“1A87CA442K”

https://stedolan.github.io/jq/

https://stedolan.github.io/jq/



Why HTTPS, REST, and JSON?

HTTPS: The Web Protocol Well-understood by admins Known security model Known network configuration

REST: The API Architecture Has largely replaced SOAP for most web APIs

JSON: The Data Format Human-readable Simpler than XML Cross language support

The combination of language support and ubiquity of REST, HTTP and JSON means that systems management tasks can be performed using the same skill set and tool chain as all other IT and dev/ops tasks.

9



Slightly more complex example

Using the same Computer System as the previous example, but this time we want to physically locate it in the data center. To do this we need to start the IndicatorLED on the computer system to “Blinking”.

Same paradigm for indicating a locally attached disk drive as well

10

echo {\"IndicatorLED\":\"Blinking\"} > led.json curl -X PATCH --data “@led.json" -u login:password http:// 192.168.1.135/redfish/v1/Systems/1

echo {\"IndicatorLED\":\"Blinking\"} > led.json curl -X PATCH --data “@led.json" -u login:password http:// 192.168.1.135/redfish/v1/Systems/1/Storage/1/Drives/1



Discovery

The only well known URI for a Redfish Service is /redfish/ The contents should look like this:

{ "v1": "/redfish/v1/" }

This indicates that this Redfish Service supports v1 at the URI /redfish/v1

11



Discovery

Once the user has obtained the URI for the version they want, the user can obtain the Service Root. The contents should something look like this:

{ "@odata.context": "http://192.168.1.135/redfish/v1/$metadata#RootService", "@odata.id": "/redfish/v1/", "@odata.type": "#ServiceRoot.v1_1_0.ServiceRoot", "Id": "RootService", "Name":"Root Service", "RedfishVersion":"1.0.4", "UUID":"92384634-2938-2342-8820-489239905423", "Chassis": {"@odata.id":"/redfish/v1/Chassis"}, "Mangers": {"@odata.id":"/redfish/v1/Managers"}, “Systems": {"@odata.id":"/redfish/v1/Systems"} }

12



Discovery The service root contains the root level entries for a Redfish Service.

Currently this may include: Systems – A collection of all Application Servers Chassis – A collection of all the Chassis Managers – A collection of all the Managers Task – A task management service SessionService – A session management service AccountService – An account management service EventService – An event management and subscription service Registries – A set of message registries JsonSchemas – A set of JSON schema files describing the

entities supported by this service Fabrics – A collection of simple fabric models UpdateService – A firmware update service

13



Discovery Each of those entity types can then contain more entity types. For

example ComputerSystem (under Systems from the Service Root) can contain: Processors – A collection of all processors in the computer

system EthernetInterfaces – A collection of all Ethernet network

interfaces SimpleStorage – A collection of all very simple HBA style

storage devices LogServices – A collection of log services Storage – A collection of more complex storage devices

14



Discovery So let’s take an example of discovering every Drive modeled by a

Redfish Service. Drives are available under Storage which is under ComputerSystem. So first we discover all ComputerSystems in our Redfish

Service. From our ServiceRoot the URI was “/redfish/v1/Systems”

15

curl http://192.168.1.135/redfish/v1/Systems

{ "@odata.context": "/redfish/v1/$metadata#ComputerSystemCollection", "@odata.type": "#ComputerSystemCollection.ComputerSystemCollection", "Name": “Systems Collection", "[email protected]": 1, "Members": [ { "@odata.id": "/redfish/v1/Systems/1" } ], "@odata.id": "/redfish/v1/Systems" }



Discovery In this example we have 1 ComputerSystem located at

"/redfish/v1/Systems/1". Now we need to know the URI for the StorageCollection of all

Storage Devices hosted by this ComputerSystem

In this example, and in fact quite often, the URI’s are simply the name of the node added to the end of the parent URI, this is not guaranteed.

16

curl http://192.168.1.135/redfish/v1/Systems/1 | jq “.Storage”

{ "@odata.id": "/redfish/v1/Systems/1/Storage" }



Discovery Now we need to know how many storage entities are available on

this computer system.

17

curl http://192.168.1.135/redfish/v1/Systems/1/Storage

{ "@odata.context": "/redfish/v1/$metadata#StorageCollection", "@odata.type": "#StorageCollection.StorageCollection", "Name": “Storage Collection", "[email protected]": 2, "Members": [ { "@odata.id": "/redfish/v1/Systems/1/Storage/SATA" }, { "@odata.id": "/redfish/v1/Systems/1/Storage/RAID" } ], "@odata.id": "/redfish/v1/Systems/1/Storage" }



Discovery Now the user could query each Storage Entity determine the URI to

each Drive entity and then query that, but there is a short cut.

18

curl http://192.168.1.135/redfish/v1/Systems/1/Storage/SATA?$expand=Drives | jq “.Drives”

[ { "Id": "0", "Name": "Physical Disk 0", "IndicatorLED": "Lit", "Model": “12345”, "Revision":"S20A", "CapacityBytes": 899527000000, … }, { "Id": “1", "Name": "Physical Disk 1", "IndicatorLED": "Lit", "Model": “12345”, "Revision":"S20A", "CapacityBytes": 899527000000, … }, … ]



Creation Example Obtaining data from the API and setting simple attributes is useful,

but sometimes you need to do something more complex. Here is an example of how to create a Volume (Virtual Disk/Logical Disk) on an example system.

This example just created a Volume named “Test Drive” with a size of 1TiB utilizing a mirror for redundancy.

19

echo {\“Name\":\“Test Drive\“, \“CapacityBytes\”: 1099511627776, \“VolumeType\”: \“Mirrored\”} > volume.json curl -X POST --data “@volume.json" -u login:password http:// 192.168.1.135/redfish/v1/Systems/1/Storage/RAID/Volumes



Backup

20



Delete Volume

21

curl -X DELETE -u login:password http:// 192.168.1.135/redfish/v1/Systems/1/Storage/RAID/Volumes/1



First we need to give the Storage Node an Encryption Key to work with

Now Encrypt the Volume

Enable Encryption and Encrypt existing Volume

22

echo {\“EncryptionKey\":\“Test Key\“} > key.json curl -X POST --data “@key.json" -u login:password http:// 192.168.1.135/redfish/v1/Systems/1/Storage/RAID/Actions/Storage.SetEncryptionKey

echo {\“Encrypted\": true} > encrypt.json curl -X PATCH --data “@encrypt.json" -u login:password http:// 192.168.1.135/redfish/v1/Systems/1/Storage/RAID/Volumes/1



In our initial example we did not specify which drives to use in the Volume. This allows the Redfish Service to choose the drives. However, that may not be the drives desired by the client. Here is how to specify drives in the Volume creation.

Create Volume with Drives specified

23

echo {\“Name\":\“Test Drive\“, \“CapacityBytes\”: 1099511627776, \“VolumeType\”: \“Mirrored\”, \“Links\”: {\“Drives\”: [{\”@odata.id\”: \”/redfish/v1/Systems/1/Storage/RAID/Drives/1\”}, {\”@odata.id\”: \”/redfish/v1/Systems/1/Storage/RAID/Drives/2\”}]}} > volume.json curl -X POST --data “@volume.json" -u login:password http:// 192.168.1.135/redfish/v1/Systems/1/Storage/RAID/Volumes

introduction and overview of redfish - snia | … fujitsu, huawei, ibm, insyde software, mellanox,...

Documents